Provider News:
UpCloud will introduce new lines of Cloud Servers in May
Instance Type: NVIDIA HGX H200
This plan belongs to the DigitalOcean NVIDIA HGX H200 server line (GPU Instances).
Instance Type Description
The NVIDIA HGX H200 Droplet is a state-of-the-art GPU instance designed for high-performance AI workloads, including large language model training and real-time inference. Built on the Hopper architecture, it features 141GB of HBM3e memory and 4.8 TB/s of bandwidth, delivering up to 2x faster inference speeds than the H100.
Instance Type Specifications
| Hourly billing | yes |
| CPU type | NVIDIA |
| Shared CPU | no |
| Storage included | yes |
See also all DigitalOcean instance types.
"NVIDIA H200" Plan Specifications
| Provider | DigitalOcean |
|---|---|
| Plan URL | https://www.digitalocean.com/pricing/gpu-droplets#on-demand-plans |
| Price per hour | $3.44 |
| GPU count and Model | 1 x NVIDIA H200 |
| vCPUs | 24 |
| RAM | 240.0 GB |
| Disk GB | 5720 |
| Data Transfer GB | 15000 |